[deliverable/linux.git] / net / core / net_namespace.c

#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

#include <linux/workqueue.h>
#include <linux/rtnetlink.h>
#include <linux/cache.h>
#include <linux/slab.h>
#include <linux/list.h>
#include <linux/delay.h>
#include <linux/sched.h>
#include <linux/idr.h>
#include <linux/rculist.h>
#include <linux/nsproxy.h>
#include <linux/proc_fs.h>
#include <linux/file.h>
#include <linux/export.h>
#include <net/net_namespace.h>
#include <net/netns/generic.h>

/*
 *	Our network namespace constructor/destructor lists
 */

static LIST_HEAD(pernet_list);
static struct list_head *first_device = &pernet_list;
static DEFINE_MUTEX(net_mutex);

LIST_HEAD(net_namespace_list);
EXPORT_SYMBOL_GPL(net_namespace_list);

struct net init_net = {
	.dev_base_head = LIST_HEAD_INIT(init_net.dev_base_head),
};
EXPORT_SYMBOL(init_net);

#define INITIAL_NET_GEN_PTRS	13 /* +1 for len +2 for rcu_head */

static unsigned int max_gen_ptrs = INITIAL_NET_GEN_PTRS;

static struct net_generic *net_alloc_generic(void)
{
	struct net_generic *ng;
	size_t generic_size = offsetof(struct net_generic, ptr[max_gen_ptrs]);

	ng = kzalloc(generic_size, GFP_KERNEL);
	if (ng)
		ng->len = max_gen_ptrs;

	return ng;
}

static int net_assign_generic(struct net *net, int id, void *data)
{
	struct net_generic *ng, *old_ng;

	BUG_ON(!mutex_is_locked(&net_mutex));
	BUG_ON(id == 0);

	old_ng = rcu_dereference_protected(net->gen,
					   lockdep_is_held(&net_mutex));
	ng = old_ng;
	if (old_ng->len >= id)
		goto assign;

	ng = net_alloc_generic();
	if (ng == NULL)
		return -ENOMEM;

	/*
	 * Some synchronisation notes:
	 *
	 * The net_generic explores the net->gen array inside rcu
	 * read section. Besides once set the net->gen->ptr[x]
	 * pointer never changes (see rules in netns/generic.h).
	 *
	 * That said, we simply duplicate this array and schedule
	 * the old copy for kfree after a grace period.
	 */

	memcpy(&ng->ptr, &old_ng->ptr, old_ng->len * sizeof(void*));

	rcu_assign_pointer(net->gen, ng);
	kfree_rcu(old_ng, rcu);
assign:
	ng->ptr[id - 1] = data;
	return 0;
}

static int ops_init(const struct pernet_operations *ops, struct net *net)
{
	int err = -ENOMEM;
	void *data = NULL;

	if (ops->id && ops->size) {
		data = kzalloc(ops->size, GFP_KERNEL);
		if (!data)
			goto out;

		err = net_assign_generic(net, *ops->id, data);
		if (err)
			goto cleanup;
	}
	err = 0;
	if (ops->init)
		err = ops->init(net);
	if (!err)
		return 0;

cleanup:
	kfree(data);

out:
	return err;
}

static void ops_free(const struct pernet_operations *ops, struct net *net)
{
	if (ops->id && ops->size) {
		int id = *ops->id;
		kfree(net_generic(net, id));
	}
}

static void ops_exit_list(const struct pernet_operations *ops,
			  struct list_head *net_exit_list)
{
	struct net *net;
	if (ops->exit) {
		list_for_each_entry(net, net_exit_list, exit_list)
			ops->exit(net);
	}
	if (ops->exit_batch)
		ops->exit_batch(net_exit_list);
}

static void ops_free_list(const struct pernet_operations *ops,
			  struct list_head *net_exit_list)
{
	struct net *net;
	if (ops->size && ops->id) {
		list_for_each_entry(net, net_exit_list, exit_list)
			ops_free(ops, net);
	}
}

/*
 * setup_net runs the initializers for the network namespace object.
 */
static __net_init int setup_net(struct net *net)
{
	/* Must be called with net_mutex held */
	const struct pernet_operations *ops, *saved_ops;
	int error = 0;
	LIST_HEAD(net_exit_list);

	atomic_set(&net->count, 1);
	atomic_set(&net->passive, 1);
	net->dev_base_seq = 1;

#ifdef NETNS_REFCNT_DEBUG
	atomic_set(&net->use_count, 0);
#endif

	list_for_each_entry(ops, &pernet_list, list) {
		error = ops_init(ops, net);
		if (error < 0)
			goto out_undo;
	}
out:
	return error;

out_undo:
	/* Walk through the list backwards calling the exit functions
	 * for the pernet modules whose init functions did not fail.
	 */
	list_add(&net->exit_list, &net_exit_list);
	saved_ops = ops;
	list_for_each_entry_continue_reverse(ops, &pernet_list, list)
		ops_exit_list(ops, &net_exit_list);

	ops = saved_ops;
	list_for_each_entry_continue_reverse(ops, &pernet_list, list)
		ops_free_list(ops, &net_exit_list);

	rcu_barrier();
	goto out;
}


#ifdef CONFIG_NET_NS
static struct kmem_cache *net_cachep;
static struct workqueue_struct *netns_wq;

static struct net *net_alloc(void)
{
	struct net *net = NULL;
	struct net_generic *ng;

	ng = net_alloc_generic();
	if (!ng)
		goto out;

	net = kmem_cache_zalloc(net_cachep, GFP_KERNEL);
	if (!net)
		goto out_free;

	rcu_assign_pointer(net->gen, ng);
out:
	return net;

out_free:
	kfree(ng);
	goto out;
}

static void net_free(struct net *net)
{
#ifdef NETNS_REFCNT_DEBUG
	if (unlikely(atomic_read(&net->use_count) != 0)) {
		pr_emerg("network namespace not free! Usage: %d\n",
			 atomic_read(&net->use_count));
		return;
	}
#endif
	kfree(net->gen);
	kmem_cache_free(net_cachep, net);
}

void net_drop_ns(void *p)
{
	struct net *ns = p;
	if (ns && atomic_dec_and_test(&ns->passive))
		net_free(ns);
}

struct net *copy_net_ns(unsigned long flags, struct net *old_net)
{
	struct net *net;
	int rv;

	if (!(flags & CLONE_NEWNET))
		return get_net(old_net);

	net = net_alloc();
	if (!net)
		return ERR_PTR(-ENOMEM);
	mutex_lock(&net_mutex);
	rv = setup_net(net);
	if (rv == 0) {
		rtnl_lock();
		list_add_tail_rcu(&net->list, &net_namespace_list);
		rtnl_unlock();
	}
	mutex_unlock(&net_mutex);
	if (rv < 0) {
		net_drop_ns(net);
		return ERR_PTR(rv);
	}
	return net;
}

static DEFINE_SPINLOCK(cleanup_list_lock);
static LIST_HEAD(cleanup_list);  /* Must hold cleanup_list_lock to touch */

static void cleanup_net(struct work_struct *work)
{
	const struct pernet_operations *ops;
	struct net *net, *tmp;
	LIST_HEAD(net_kill_list);
	LIST_HEAD(net_exit_list);

	/* Atomically snapshot the list of namespaces to cleanup */
	spin_lock_irq(&cleanup_list_lock);
	list_replace_init(&cleanup_list, &net_kill_list);
	spin_unlock_irq(&cleanup_list_lock);

	mutex_lock(&net_mutex);

	/* Don't let anyone else find us. */
	rtnl_lock();
	list_for_each_entry(net, &net_kill_list, cleanup_list) {
		list_del_rcu(&net->list);
		list_add_tail(&net->exit_list, &net_exit_list);
	}
	rtnl_unlock();

	/*
	 * Another CPU might be rcu-iterating the list, wait for it.
	 * This needs to be before calling the exit() notifiers, so
	 * the rcu_barrier() below isn't sufficient alone.
	 */
	synchronize_rcu();

	/* Run all of the network namespace exit methods */
	list_for_each_entry_reverse(ops, &pernet_list, list)
		ops_exit_list(ops, &net_exit_list);

	/* Free the net generic variables */
	list_for_each_entry_reverse(ops, &pernet_list, list)
		ops_free_list(ops, &net_exit_list);

	mutex_unlock(&net_mutex);

	/* Ensure there are no outstanding rcu callbacks using this
	 * network namespace.
	 */
	rcu_barrier();

	/* Finally it is safe to free my network namespace structure */
	list_for_each_entry_safe(net, tmp, &net_exit_list, exit_list) {
		list_del_init(&net->exit_list);
		net_drop_ns(net);
	}
}
static DECLARE_WORK(net_cleanup_work, cleanup_net);

void __put_net(struct net *net)
{
	/* Cleanup the network namespace in process context */
	unsigned long flags;

	spin_lock_irqsave(&cleanup_list_lock, flags);
	list_add(&net->cleanup_list, &cleanup_list);
	spin_unlock_irqrestore(&cleanup_list_lock, flags);

	queue_work(netns_wq, &net_cleanup_work);
}
EXPORT_SYMBOL_GPL(__put_net);

struct net *get_net_ns_by_fd(int fd)
{
	struct proc_inode *ei;
	struct file *file;
	struct net *net;

	file = proc_ns_fget(fd);
	if (IS_ERR(file))
		return ERR_CAST(file);

	ei = PROC_I(file->f_dentry->d_inode);
	if (ei->ns_ops == &netns_operations)
		net = get_net(ei->ns);
	else
		net = ERR_PTR(-EINVAL);

	fput(file);
	return net;
}

#else
struct net *copy_net_ns(unsigned long flags, struct net *old_net)
{
	if (flags & CLONE_NEWNET)
		return ERR_PTR(-EINVAL);
	return old_net;
}

struct net *get_net_ns_by_fd(int fd)
{
	return ERR_PTR(-EINVAL);
}
#endif

struct net *get_net_ns_by_pid(pid_t pid)
{
	struct task_struct *tsk;
	struct net *net;

	/* Lookup the network namespace */
	net = ERR_PTR(-ESRCH);
	rcu_read_lock();
	tsk = find_task_by_vpid(pid);
	if (tsk) {
		struct nsproxy *nsproxy;
		nsproxy = task_nsproxy(tsk);
		if (nsproxy)
			net = get_net(nsproxy->net_ns);
	}
	rcu_read_unlock();
	return net;
}
EXPORT_SYMBOL_GPL(get_net_ns_by_pid);

static int __init net_ns_init(void)
{
	struct net_generic *ng;

#ifdef CONFIG_NET_NS
	net_cachep = kmem_cache_create("net_namespace", sizeof(struct net),
					SMP_CACHE_BYTES,
					SLAB_PANIC, NULL);

	/* Create workqueue for cleanup */
	netns_wq = create_singlethread_workqueue("netns");
	if (!netns_wq)
		panic("Could not create netns workq");
#endif

	ng = net_alloc_generic();
	if (!ng)
		panic("Could not allocate generic netns");

	rcu_assign_pointer(init_net.gen, ng);

	mutex_lock(&net_mutex);
	if (setup_net(&init_net))
		panic("Could not setup the initial network namespace");

	rtnl_lock();
	list_add_tail_rcu(&init_net.list, &net_namespace_list);
	rtnl_unlock();

	mutex_unlock(&net_mutex);

	return 0;
}

pure_initcall(net_ns_init);

#ifdef CONFIG_NET_NS
static int __register_pernet_operations(struct list_head *list,
					struct pernet_operations *ops)
{
	struct net *net;
	int error;
	LIST_HEAD(net_exit_list);

	list_add_tail(&ops->list, list);
	if (ops->init || (ops->id && ops->size)) {
		for_each_net(net) {
			error = ops_init(ops, net);
			if (error)
				goto out_undo;
			list_add_tail(&net->exit_list, &net_exit_list);
		}
	}
	return 0;

out_undo:
	/* If I have an error cleanup all namespaces I initialized */
	list_del(&ops->list);
	ops_exit_list(ops, &net_exit_list);
	ops_free_list(ops, &net_exit_list);
	return error;
}

static void __unregister_pernet_operations(struct pernet_operations *ops)
{
	struct net *net;
	LIST_HEAD(net_exit_list);

	list_del(&ops->list);
	for_each_net(net)
		list_add_tail(&net->exit_list, &net_exit_list);
	ops_exit_list(ops, &net_exit_list);
	ops_free_list(ops, &net_exit_list);
}

#else

static int __register_pernet_operations(struct list_head *list,
					struct pernet_operations *ops)
{
	return ops_init(ops, &init_net);
}

static void __unregister_pernet_operations(struct pernet_operations *ops)
{
	LIST_HEAD(net_exit_list);
	list_add(&init_net.exit_list, &net_exit_list);
	ops_exit_list(ops, &net_exit_list);
	ops_free_list(ops, &net_exit_list);
}

#endif /* CONFIG_NET_NS */

static DEFINE_IDA(net_generic_ids);

static int register_pernet_operations(struct list_head *list,
				      struct pernet_operations *ops)
{
	int error;

	if (ops->id) {
again:
		error = ida_get_new_above(&net_generic_ids, 1, ops->id);
		if (error < 0) {
			if (error == -EAGAIN) {
				ida_pre_get(&net_generic_ids, GFP_KERNEL);
				goto again;
			}
			return error;
		}
		max_gen_ptrs = max_t(unsigned int, max_gen_ptrs, *ops->id);
	}
	error = __register_pernet_operations(list, ops);
	if (error) {
		rcu_barrier();
		if (ops->id)
			ida_remove(&net_generic_ids, *ops->id);
	}

	return error;
}

static void unregister_pernet_operations(struct pernet_operations *ops)
{
	
	__unregister_pernet_operations(ops);
	rcu_barrier();
	if (ops->id)
		ida_remove(&net_generic_ids, *ops->id);
}

/**
 *      register_pernet_subsys - register a network namespace subsystem
 *	@ops:  pernet operations structure for the subsystem
 *
 *	Register a subsystem which has init and exit functions
 *	that are called when network namespaces are created and
 *	destroyed respectively.
 *
 *	When registered all network namespace init functions are
 *	called for every existing network namespace.  Allowing kernel
 *	modules to have a race free view of the set of network namespaces.
 *
 *	When a new network namespace is created all of the init
 *	methods are called in the order in which they were registered.
 *
 *	When a network namespace is destroyed all of the exit methods
 *	are called in the reverse of the order with which they were
 *	registered.
 */
int register_pernet_subsys(struct pernet_operations *ops)
{
	int error;
	mutex_lock(&net_mutex);
	error =  register_pernet_operations(first_device, ops);
	mutex_unlock(&net_mutex);
	return error;
}
EXPORT_SYMBOL_GPL(register_pernet_subsys);

/**
 *      unregister_pernet_subsys - unregister a network namespace subsystem
 *	@ops: pernet operations structure to manipulate
 *
 *	Remove the pernet operations structure from the list to be
 *	used when network namespaces are created or destroyed.  In
 *	addition run the exit method for all existing network
 *	namespaces.
 */
void unregister_pernet_subsys(struct pernet_operations *ops)
{
	mutex_lock(&net_mutex);
	unregister_pernet_operations(ops);
	mutex_unlock(&net_mutex);
}
EXPORT_SYMBOL_GPL(unregister_pernet_subsys);

/**
 *      register_pernet_device - register a network namespace device
 *	@ops:  pernet operations structure for the subsystem
 *
 *	Register a device which has init and exit functions
 *	that are called when network namespaces are created and
 *	destroyed respectively.
 *
 *	When registered all network namespace init functions are
 *	called for every existing network namespace.  Allowing kernel
 *	modules to have a race free view of the set of network namespaces.
 *
 *	When a new network namespace is created all of the init
 *	methods are called in the order in which they were registered.
 *
 *	When a network namespace is destroyed all of the exit methods
 *	are called in the reverse of the order with which they were
 *	registered.
 */
int register_pernet_device(struct pernet_operations *ops)
{
	int error;
	mutex_lock(&net_mutex);
	error = register_pernet_operations(&pernet_list, ops);
	if (!error && (first_device == &pernet_list))
		first_device = &ops->list;
	mutex_unlock(&net_mutex);
	return error;
}
EXPORT_SYMBOL_GPL(register_pernet_device);

/**
 *      unregister_pernet_device - unregister a network namespace netdevice
 *	@ops: pernet operations structure to manipulate
 *
 *	Remove the pernet operations structure from the list to be
 *	used when network namespaces are created or destroyed.  In
 *	addition run the exit method for all existing network
 *	namespaces.
 */
void unregister_pernet_device(struct pernet_operations *ops)
{
	mutex_lock(&net_mutex);
	if (&ops->list == first_device)
		first_device = first_device->next;
	unregister_pernet_operations(ops);
	mutex_unlock(&net_mutex);
}
EXPORT_SYMBOL_GPL(unregister_pernet_device);

#ifdef CONFIG_NET_NS
static void *netns_get(struct task_struct *task)
{
	struct net *net = NULL;
	struct nsproxy *nsproxy;

	rcu_read_lock();
	nsproxy = task_nsproxy(task);
	if (nsproxy)
		net = get_net(nsproxy->net_ns);
	rcu_read_unlock();

	return net;
}

static void netns_put(void *ns)
{
	put_net(ns);
}

static int netns_install(struct nsproxy *nsproxy, void *ns)
{
	put_net(nsproxy->net_ns);
	nsproxy->net_ns = get_net(ns);
	return 0;
}

const struct proc_ns_operations netns_operations = {
	.name		= "net",
	.type		= CLONE_NEWNET,
	.get		= netns_get,
	.put		= netns_put,
	.install	= netns_install,
};
#endif
Commit	Line	Data
e005d193 JP	1	#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
e005d193 JP	2
5f256bec EB	3	#include <linux/workqueue.h>
	4	#include <linux/rtnetlink.h>
	5	#include <linux/cache.h>
	6	#include <linux/slab.h>
	7	#include <linux/list.h>
	8	#include <linux/delay.h>
9dd776b6	9	#include <linux/sched.h>
c93cf61f	10	#include <linux/idr.h>
11a28d37	11	#include <linux/rculist.h>
30ffee84	12	#include <linux/nsproxy.h>
f0630529 EB	13	#include <linux/proc_fs.h>
f0630529 EB	14	#include <linux/file.h>
bc3b2d7f	15	#include <linux/export.h>
5f256bec	16	#include <net/net_namespace.h>
dec827d1	17	#include <net/netns/generic.h>
5f256bec EB	18
	19	/*
	20	* Our network namespace constructor/destructor lists
	21	*/
	22
	23	static LIST_HEAD(pernet_list);
	24	static struct list_head *first_device = &pernet_list;
	25	static DEFINE_MUTEX(net_mutex);
	26
5f256bec	27	LIST_HEAD(net_namespace_list);
b76a461f	28	EXPORT_SYMBOL_GPL(net_namespace_list);
5f256bec	29
734b6541 RM	30	struct net init_net = {
	31	.dev_base_head = LIST_HEAD_INIT(init_net.dev_base_head),
	32	};
ff4b9502	33	EXPORT_SYMBOL(init_net);
5f256bec	34
dec827d1 PE	35	#define INITIAL_NET_GEN_PTRS 13 /* +1 for len +2 for rcu_head */
dec827d1 PE	36
073862ba ED	37	static unsigned int max_gen_ptrs = INITIAL_NET_GEN_PTRS;
	38
	39	static struct net_generic *net_alloc_generic(void)
	40	{
	41	struct net_generic *ng;
	42	size_t generic_size = offsetof(struct net_generic, ptr[max_gen_ptrs]);
	43
	44	ng = kzalloc(generic_size, GFP_KERNEL);
	45	if (ng)
	46	ng->len = max_gen_ptrs;
	47
	48	return ng;
	49	}
	50
05fceb4a JP	51	static int net_assign_generic(struct net net, int id, void data)
	52	{
	53	struct net_generic ng, old_ng;
	54
	55	BUG_ON(!mutex_is_locked(&net_mutex));
	56	BUG_ON(id == 0);
	57
1c87733d ED	58	old_ng = rcu_dereference_protected(net->gen,
	59	lockdep_is_held(&net_mutex));
	60	ng = old_ng;
05fceb4a JP	61	if (old_ng->len >= id)
	62	goto assign;
	63
073862ba	64	ng = net_alloc_generic();
05fceb4a JP	65	if (ng == NULL)
	66	return -ENOMEM;
	67
	68	/*
	69	* Some synchronisation notes:
	70	*
	71	* The net_generic explores the net->gen array inside rcu
	72	* read section. Besides once set the net->gen->ptr[x]
	73	* pointer never changes (see rules in netns/generic.h).
	74	*
	75	* That said, we simply duplicate this array and schedule
	76	* the old copy for kfree after a grace period.
	77	*/
	78
05fceb4a JP	79	memcpy(&ng->ptr, &old_ng->ptr, old_ng->len * sizeof(void*));
	80
	81	rcu_assign_pointer(net->gen, ng);
04d4dfed	82	kfree_rcu(old_ng, rcu);
05fceb4a JP	83	assign:
	84	ng->ptr[id - 1] = data;
	85	return 0;
	86	}
	87
f875bae0 EB	88	static int ops_init(const struct pernet_operations ops, struct net net)
f875bae0 EB	89	{
b922934d JA	90	int err = -ENOMEM;
	91	void *data = NULL;
	92
f875bae0	93	if (ops->id && ops->size) {
b922934d	94	data = kzalloc(ops->size, GFP_KERNEL);
f875bae0	95	if (!data)
b922934d	96	goto out;
f875bae0 EB	97
f875bae0 EB	98	err = net_assign_generic(net, *ops->id, data);
b922934d JA	99	if (err)
b922934d JA	100	goto cleanup;
f875bae0	101	}
b922934d	102	err = 0;
f875bae0	103	if (ops->init)
b922934d JA	104	err = ops->init(net);
	105	if (!err)
	106	return 0;
	107
	108	cleanup:
	109	kfree(data);
	110
	111	out:
	112	return err;
f875bae0 EB	113	}
	114
	115	static void ops_free(const struct pernet_operations ops, struct net net)
	116	{
	117	if (ops->id && ops->size) {
	118	int id = *ops->id;
	119	kfree(net_generic(net, id));
	120	}
	121	}
	122
72ad937a EB	123	static void ops_exit_list(const struct pernet_operations *ops,
	124	struct list_head *net_exit_list)
	125	{
	126	struct net *net;
	127	if (ops->exit) {
	128	list_for_each_entry(net, net_exit_list, exit_list)
	129	ops->exit(net);
	130	}
72ad937a EB	131	if (ops->exit_batch)
	132	ops->exit_batch(net_exit_list);
	133	}
	134
	135	static void ops_free_list(const struct pernet_operations *ops,
	136	struct list_head *net_exit_list)
	137	{
	138	struct net *net;
	139	if (ops->size && ops->id) {
	140	list_for_each_entry(net, net_exit_list, exit_list)
	141	ops_free(ops, net);
	142	}
	143	}
	144
5f256bec EB	145	/*
	146	* setup_net runs the initializers for the network namespace object.
	147	*/
1a2ee93d	148	static __net_init int setup_net(struct net *net)
5f256bec EB	149	{
5f256bec EB	150	/* Must be called with net_mutex held */
f875bae0	151	const struct pernet_operations ops, saved_ops;
486a87f1	152	int error = 0;
72ad937a	153	LIST_HEAD(net_exit_list);
5f256bec	154
5f256bec	155	atomic_set(&net->count, 1);
a685e089	156	atomic_set(&net->passive, 1);
4e985ada	157	net->dev_base_seq = 1;
486a87f1	158
5d1e4468	159	#ifdef NETNS_REFCNT_DEBUG
5f256bec	160	atomic_set(&net->use_count, 0);
5d1e4468	161	#endif
5f256bec	162
768f3591	163	list_for_each_entry(ops, &pernet_list, list) {
f875bae0 EB	164	error = ops_init(ops, net);
	165	if (error < 0)
	166	goto out_undo;
5f256bec EB	167	}
	168	out:
	169	return error;
768f3591	170
5f256bec EB	171	out_undo:
	172	/* Walk through the list backwards calling the exit functions
	173	* for the pernet modules whose init functions did not fail.
	174	*/
72ad937a	175	list_add(&net->exit_list, &net_exit_list);
f875bae0	176	saved_ops = ops;
72ad937a EB	177	list_for_each_entry_continue_reverse(ops, &pernet_list, list)
	178	ops_exit_list(ops, &net_exit_list);
	179
f875bae0 EB	180	ops = saved_ops;
f875bae0 EB	181	list_for_each_entry_continue_reverse(ops, &pernet_list, list)
72ad937a	182	ops_free_list(ops, &net_exit_list);
310928d9 DL	183
310928d9 DL	184	rcu_barrier();
5f256bec EB	185	goto out;
	186	}
	187
6a1a3b9f	188
ebe47d47 CN	189	#ifdef CONFIG_NET_NS
	190	static struct kmem_cache *net_cachep;
	191	static struct workqueue_struct *netns_wq;
	192
486a87f1	193	static struct net *net_alloc(void)
45a19b0a	194	{
486a87f1 DL	195	struct net *net = NULL;
	196	struct net_generic *ng;
	197
	198	ng = net_alloc_generic();
	199	if (!ng)
	200	goto out;
	201
	202	net = kmem_cache_zalloc(net_cachep, GFP_KERNEL);
45a19b0a	203	if (!net)
486a87f1	204	goto out_free;
45a19b0a	205
486a87f1 DL	206	rcu_assign_pointer(net->gen, ng);
	207	out:
	208	return net;
	209
	210	out_free:
	211	kfree(ng);
	212	goto out;
	213	}
	214
	215	static void net_free(struct net *net)
	216	{
5d1e4468	217	#ifdef NETNS_REFCNT_DEBUG
45a19b0a	218	if (unlikely(atomic_read(&net->use_count) != 0)) {
e005d193 JP	219	pr_emerg("network namespace not free! Usage: %d\n",
e005d193 JP	220	atomic_read(&net->use_count));
45a19b0a JFS	221	return;
45a19b0a JFS	222	}
5d1e4468	223	#endif
4ef079cc	224	kfree(net->gen);
45a19b0a JFS	225	kmem_cache_free(net_cachep, net);
	226	}
	227
a685e089 AV	228	void net_drop_ns(void *p)
	229	{
	230	struct net *ns = p;
	231	if (ns && atomic_dec_and_test(&ns->passive))
	232	net_free(ns);
	233	}
	234
911cb193	235	struct net copy_net_ns(unsigned long flags, struct net old_net)
9dd776b6	236	{
088eb2d9 AD	237	struct net *net;
088eb2d9 AD	238	int rv;
9dd776b6	239
911cb193 RL	240	if (!(flags & CLONE_NEWNET))
	241	return get_net(old_net);
	242
088eb2d9 AD	243	net = net_alloc();
	244	if (!net)
	245	return ERR_PTR(-ENOMEM);
9dd776b6	246	mutex_lock(&net_mutex);
088eb2d9 AD	247	rv = setup_net(net);
088eb2d9 AD	248	if (rv == 0) {
486a87f1	249	rtnl_lock();
11a28d37	250	list_add_tail_rcu(&net->list, &net_namespace_list);
486a87f1 DL	251	rtnl_unlock();
486a87f1 DL	252	}
9dd776b6	253	mutex_unlock(&net_mutex);
088eb2d9	254	if (rv < 0) {
a685e089	255	net_drop_ns(net);
088eb2d9 AD	256	return ERR_PTR(rv);
	257	}
	258	return net;
	259	}
486a87f1	260
2b035b39 EB	261	static DEFINE_SPINLOCK(cleanup_list_lock);
	262	static LIST_HEAD(cleanup_list); /* Must hold cleanup_list_lock to touch */
	263
6a1a3b9f PE	264	static void cleanup_net(struct work_struct *work)
6a1a3b9f PE	265	{
f875bae0	266	const struct pernet_operations *ops;
2b035b39 EB	267	struct net net, tmp;
2b035b39 EB	268	LIST_HEAD(net_kill_list);
72ad937a	269	LIST_HEAD(net_exit_list);
6a1a3b9f	270
2b035b39 EB	271	/* Atomically snapshot the list of namespaces to cleanup */
	272	spin_lock_irq(&cleanup_list_lock);
	273	list_replace_init(&cleanup_list, &net_kill_list);
	274	spin_unlock_irq(&cleanup_list_lock);
6a1a3b9f PE	275
	276	mutex_lock(&net_mutex);
	277
	278	/* Don't let anyone else find us. */
	279	rtnl_lock();
72ad937a	280	list_for_each_entry(net, &net_kill_list, cleanup_list) {
2b035b39	281	list_del_rcu(&net->list);
72ad937a EB	282	list_add_tail(&net->exit_list, &net_exit_list);
72ad937a EB	283	}
6a1a3b9f PE	284	rtnl_unlock();
6a1a3b9f PE	285
11a28d37 JB	286	/*
	287	* Another CPU might be rcu-iterating the list, wait for it.
	288	* This needs to be before calling the exit() notifiers, so
	289	* the rcu_barrier() below isn't sufficient alone.
	290	*/
	291	synchronize_rcu();
	292
6a1a3b9f	293	/* Run all of the network namespace exit methods */
72ad937a EB	294	list_for_each_entry_reverse(ops, &pernet_list, list)
	295	ops_exit_list(ops, &net_exit_list);
	296
f875bae0	297	/* Free the net generic variables */
72ad937a EB	298	list_for_each_entry_reverse(ops, &pernet_list, list)
72ad937a EB	299	ops_free_list(ops, &net_exit_list);
6a1a3b9f PE	300
	301	mutex_unlock(&net_mutex);
	302
	303	/* Ensure there are no outstanding rcu callbacks using this
	304	* network namespace.
	305	*/
	306	rcu_barrier();
	307
	308	/* Finally it is safe to free my network namespace structure */
72ad937a EB	309	list_for_each_entry_safe(net, tmp, &net_exit_list, exit_list) {
72ad937a EB	310	list_del_init(&net->exit_list);
a685e089	311	net_drop_ns(net);
2b035b39	312	}
6a1a3b9f	313	}
2b035b39	314	static DECLARE_WORK(net_cleanup_work, cleanup_net);
6a1a3b9f PE	315
	316	void __put_net(struct net *net)
	317	{
	318	/* Cleanup the network namespace in process context */
2b035b39 EB	319	unsigned long flags;
	320
	321	spin_lock_irqsave(&cleanup_list_lock, flags);
	322	list_add(&net->cleanup_list, &cleanup_list);
	323	spin_unlock_irqrestore(&cleanup_list_lock, flags);
	324
	325	queue_work(netns_wq, &net_cleanup_work);
6a1a3b9f PE	326	}
	327	EXPORT_SYMBOL_GPL(__put_net);
	328
956c9207 SR	329	struct net *get_net_ns_by_fd(int fd)
	330	{
	331	struct proc_inode *ei;
	332	struct file *file;
	333	struct net *net;
	334
956c9207	335	file = proc_ns_fget(fd);
c316e6a3 AV	336	if (IS_ERR(file))
c316e6a3 AV	337	return ERR_CAST(file);
956c9207 SR	338
956c9207 SR	339	ei = PROC_I(file->f_dentry->d_inode);
c316e6a3 AV	340	if (ei->ns_ops == &netns_operations)
	341	net = get_net(ei->ns);
	342	else
	343	net = ERR_PTR(-EINVAL);
956c9207	344
c316e6a3	345	fput(file);
956c9207 SR	346	return net;
	347	}
	348
6a1a3b9f PE	349	#else
	350	struct net copy_net_ns(unsigned long flags, struct net old_net)
	351	{
	352	if (flags & CLONE_NEWNET)
	353	return ERR_PTR(-EINVAL);
	354	return old_net;
	355	}
956c9207 SR	356
	357	struct net *get_net_ns_by_fd(int fd)
	358	{
	359	return ERR_PTR(-EINVAL);
	360	}
6a1a3b9f PE	361	#endif
6a1a3b9f PE	362
30ffee84 JB	363	struct net *get_net_ns_by_pid(pid_t pid)
	364	{
	365	struct task_struct *tsk;
	366	struct net *net;
	367
	368	/* Lookup the network namespace */
	369	net = ERR_PTR(-ESRCH);
	370	rcu_read_lock();
	371	tsk = find_task_by_vpid(pid);
	372	if (tsk) {
	373	struct nsproxy *nsproxy;
	374	nsproxy = task_nsproxy(tsk);
	375	if (nsproxy)
	376	net = get_net(nsproxy->net_ns);
	377	}
	378	rcu_read_unlock();
	379	return net;
	380	}
	381	EXPORT_SYMBOL_GPL(get_net_ns_by_pid);
	382
5f256bec EB	383	static int __init net_ns_init(void)
5f256bec EB	384	{
486a87f1	385	struct net_generic *ng;
5f256bec	386
d57a9212	387	#ifdef CONFIG_NET_NS
5f256bec EB	388	net_cachep = kmem_cache_create("net_namespace", sizeof(struct net),
	389	SMP_CACHE_BYTES,
	390	SLAB_PANIC, NULL);
3ef1355d BT	391
	392	/* Create workqueue for cleanup */
	393	netns_wq = create_singlethread_workqueue("netns");
	394	if (!netns_wq)
	395	panic("Could not create netns workq");
d57a9212	396	#endif
3ef1355d	397
486a87f1 DL	398	ng = net_alloc_generic();
	399	if (!ng)
	400	panic("Could not allocate generic netns");
	401
	402	rcu_assign_pointer(init_net.gen, ng);
	403
5f256bec	404	mutex_lock(&net_mutex);
ca0f3112 SH	405	if (setup_net(&init_net))
ca0f3112 SH	406	panic("Could not setup the initial network namespace");
5f256bec	407
f4618d39	408	rtnl_lock();
11a28d37	409	list_add_tail_rcu(&init_net.list, &net_namespace_list);
f4618d39	410	rtnl_unlock();
5f256bec EB	411
5f256bec EB	412	mutex_unlock(&net_mutex);
5f256bec EB	413
	414	return 0;
	415	}
	416
	417	pure_initcall(net_ns_init);
	418
ed160e83	419	#ifdef CONFIG_NET_NS
f875bae0 EB	420	static int __register_pernet_operations(struct list_head *list,
f875bae0 EB	421	struct pernet_operations *ops)
5f256bec	422	{
72ad937a	423	struct net *net;
5f256bec	424	int error;
72ad937a	425	LIST_HEAD(net_exit_list);
5f256bec	426
5f256bec	427	list_add_tail(&ops->list, list);
f875bae0	428	if (ops->init \|\| (ops->id && ops->size)) {
1dba323b	429	for_each_net(net) {
f875bae0	430	error = ops_init(ops, net);
5f256bec EB	431	if (error)
5f256bec EB	432	goto out_undo;
72ad937a	433	list_add_tail(&net->exit_list, &net_exit_list);
5f256bec EB	434	}
5f256bec EB	435	}
1dba323b	436	return 0;
5f256bec EB	437
	438	out_undo:
	439	/* If I have an error cleanup all namespaces I initialized */
	440	list_del(&ops->list);
72ad937a EB	441	ops_exit_list(ops, &net_exit_list);
72ad937a EB	442	ops_free_list(ops, &net_exit_list);
1dba323b	443	return error;
5f256bec EB	444	}
5f256bec EB	445
f875bae0	446	static void __unregister_pernet_operations(struct pernet_operations *ops)
5f256bec EB	447	{
5f256bec EB	448	struct net *net;
72ad937a	449	LIST_HEAD(net_exit_list);
5f256bec EB	450
5f256bec EB	451	list_del(&ops->list);
72ad937a EB	452	for_each_net(net)
	453	list_add_tail(&net->exit_list, &net_exit_list);
	454	ops_exit_list(ops, &net_exit_list);
	455	ops_free_list(ops, &net_exit_list);
5f256bec EB	456	}
5f256bec EB	457
ed160e83 DL	458	#else
ed160e83 DL	459
f875bae0 EB	460	static int __register_pernet_operations(struct list_head *list,
f875bae0 EB	461	struct pernet_operations *ops)
ed160e83	462	{
b922934d	463	return ops_init(ops, &init_net);
ed160e83 DL	464	}
ed160e83 DL	465
f875bae0	466	static void __unregister_pernet_operations(struct pernet_operations *ops)
ed160e83	467	{
72ad937a EB	468	LIST_HEAD(net_exit_list);
	469	list_add(&init_net.exit_list, &net_exit_list);
	470	ops_exit_list(ops, &net_exit_list);
	471	ops_free_list(ops, &net_exit_list);
ed160e83	472	}
f875bae0 EB	473
f875bae0 EB	474	#endif /* CONFIG_NET_NS */
ed160e83	475
c93cf61f PE	476	static DEFINE_IDA(net_generic_ids);
c93cf61f PE	477
f875bae0 EB	478	static int register_pernet_operations(struct list_head *list,
	479	struct pernet_operations *ops)
	480	{
	481	int error;
	482
	483	if (ops->id) {
	484	again:
	485	error = ida_get_new_above(&net_generic_ids, 1, ops->id);
	486	if (error < 0) {
	487	if (error == -EAGAIN) {
	488	ida_pre_get(&net_generic_ids, GFP_KERNEL);
	489	goto again;
	490	}
	491	return error;
	492	}
073862ba	493	max_gen_ptrs = max_t(unsigned int, max_gen_ptrs, *ops->id);
f875bae0 EB	494	}
f875bae0 EB	495	error = __register_pernet_operations(list, ops);
3a765eda EB	496	if (error) {
	497	rcu_barrier();
	498	if (ops->id)
	499	ida_remove(&net_generic_ids, *ops->id);
	500	}
f875bae0 EB	501
	502	return error;
	503	}
	504
	505	static void unregister_pernet_operations(struct pernet_operations *ops)
	506	{
	507
	508	__unregister_pernet_operations(ops);
3a765eda	509	rcu_barrier();
f875bae0 EB	510	if (ops->id)
	511	ida_remove(&net_generic_ids, *ops->id);
	512	}
	513
5f256bec EB	514	/**
	515	* register_pernet_subsys - register a network namespace subsystem
	516	* @ops: pernet operations structure for the subsystem
	517	*
	518	* Register a subsystem which has init and exit functions
	519	* that are called when network namespaces are created and
	520	* destroyed respectively.
	521	*
	522	* When registered all network namespace init functions are
	523	* called for every existing network namespace. Allowing kernel
	524	* modules to have a race free view of the set of network namespaces.
	525	*
	526	* When a new network namespace is created all of the init
	527	* methods are called in the order in which they were registered.
	528	*
	529	* When a network namespace is destroyed all of the exit methods
	530	* are called in the reverse of the order with which they were
	531	* registered.
	532	*/
	533	int register_pernet_subsys(struct pernet_operations *ops)
	534	{
	535	int error;
	536	mutex_lock(&net_mutex);
	537	error = register_pernet_operations(first_device, ops);
	538	mutex_unlock(&net_mutex);
	539	return error;
	540	}
	541	EXPORT_SYMBOL_GPL(register_pernet_subsys);
	542
	543	/**
	544	* unregister_pernet_subsys - unregister a network namespace subsystem
	545	* @ops: pernet operations structure to manipulate
	546	*
	547	* Remove the pernet operations structure from the list to be
53379e57	548	* used when network namespaces are created or destroyed. In
5f256bec EB	549	* addition run the exit method for all existing network
	550	* namespaces.
	551	*/
b3c981d2	552	void unregister_pernet_subsys(struct pernet_operations *ops)
5f256bec EB	553	{
5f256bec EB	554	mutex_lock(&net_mutex);
b3c981d2	555	unregister_pernet_operations(ops);
5f256bec EB	556	mutex_unlock(&net_mutex);
	557	}
	558	EXPORT_SYMBOL_GPL(unregister_pernet_subsys);
	559
	560	/**
	561	* register_pernet_device - register a network namespace device
	562	* @ops: pernet operations structure for the subsystem
	563	*
	564	* Register a device which has init and exit functions
	565	* that are called when network namespaces are created and
	566	* destroyed respectively.
	567	*
	568	* When registered all network namespace init functions are
	569	* called for every existing network namespace. Allowing kernel
	570	* modules to have a race free view of the set of network namespaces.
	571	*
	572	* When a new network namespace is created all of the init
	573	* methods are called in the order in which they were registered.
	574	*
	575	* When a network namespace is destroyed all of the exit methods
	576	* are called in the reverse of the order with which they were
	577	* registered.
	578	*/
	579	int register_pernet_device(struct pernet_operations *ops)
	580	{
	581	int error;
	582	mutex_lock(&net_mutex);
	583	error = register_pernet_operations(&pernet_list, ops);
	584	if (!error && (first_device == &pernet_list))
	585	first_device = &ops->list;
	586	mutex_unlock(&net_mutex);
	587	return error;
	588	}
	589	EXPORT_SYMBOL_GPL(register_pernet_device);
	590
	591	/**
	592	* unregister_pernet_device - unregister a network namespace netdevice
	593	* @ops: pernet operations structure to manipulate
	594	*
	595	* Remove the pernet operations structure from the list to be
53379e57	596	* used when network namespaces are created or destroyed. In
5f256bec EB	597	* addition run the exit method for all existing network
	598	* namespaces.
	599	*/
	600	void unregister_pernet_device(struct pernet_operations *ops)
	601	{
	602	mutex_lock(&net_mutex);
	603	if (&ops->list == first_device)
	604	first_device = first_device->next;
	605	unregister_pernet_operations(ops);
	606	mutex_unlock(&net_mutex);
	607	}
	608	EXPORT_SYMBOL_GPL(unregister_pernet_device);
13b6f576 EB	609
	610	#ifdef CONFIG_NET_NS
	611	static void netns_get(struct task_struct task)
	612	{
f0630529 EB	613	struct net *net = NULL;
	614	struct nsproxy *nsproxy;
	615
13b6f576	616	rcu_read_lock();
f0630529 EB	617	nsproxy = task_nsproxy(task);
	618	if (nsproxy)
	619	net = get_net(nsproxy->net_ns);
13b6f576	620	rcu_read_unlock();
f0630529	621
13b6f576 EB	622	return net;
	623	}
	624
	625	static void netns_put(void *ns)
	626	{
	627	put_net(ns);
	628	}
	629
	630	static int netns_install(struct nsproxy nsproxy, void ns)
	631	{
	632	put_net(nsproxy->net_ns);
	633	nsproxy->net_ns = get_net(ns);
	634	return 0;
	635	}
	636
	637	const struct proc_ns_operations netns_operations = {
	638	.name = "net",
	639	.type = CLONE_NEWNET,
	640	.get = netns_get,
	641	.put = netns_put,
	642	.install = netns_install,
	643	};
	644	#endif